RCCD in meter box randomly tripping

[FrogParty]

For the power installation on my new build the meter tails and CU are protected by a 100mA RCCD in the meter box. From the day the site was energised this has randomly tripped, often working fine when reset. The electricians have blamed the weather, the equipment in use at the time, the earth strap etc. but on other days, sometimes even just some minutes later, all of these have worked fine. As the build reaches completion we really need this issue fixing and could do with any suggestions what could be the problem, the contractors don't seem to have a clue.

The 30mA RCDs in the CU have never blown. The RCBO can trip with only one socket on one RCD in the CU (how simple can it get). The RCBO unit has been replaced to rule out a faulty unit.

Due to the shape of the building and position of the meter relative to the CU, the meter tails are 10m SWA cable in a duct buried under the floor slab. I had thought that the RCBO was overkill protection of the concealed meter tails cable, it is more than adequately protected by the material of the house. However when I suggested that the RCBO was simply removed I was told this was not safe. The protection was nothing to do with the length of the tails or that they were concealed, but because the resistance provided by our domestic earth rod means that a short in the meter tails could cause 2A to flow in the metal CU while the supply fuse would not trip.

So what is going wrong, and what can be done about it? Are the long tails somehow picking up noise and tripping the sensitive RCBO? How can we get our electrics working reliably? Anyone met anything like this before?

 

[PrenticeBoyofDerry]

If the CU is metallic on TT system as understand it you will require the 100mA RCD on the supply cable to the unit.
The reason for this tripping could be many a varied, I would suggest that the first thing to do, would an IR test (insulation resistance) on the distribution cable between RCBO and CU.

 

[flameport]

If the RCDs in the consumer unit do not trip, there is a fault with the SWA cable under the floor. Either damaged somewhere along the length, or one of the terminations is defective, perhaps one core damaged at the gland.

 

[FrogParty]

I would suggest that the first thing to do, would an IR test (insulation resistance) on the distribution cable between RCBO and CU.

Will ask for that, thanks.

If the RCDs in the consumer unit do not trip, there is a fault with the SWA cable under the floor. Either damaged somewhere along the length, or one of the terminations is defective, perhaps one core damaged at the gland.

There were difficulties threading the SWA cable into the 110mm ducting, along with a phone cable and a 25mm alkathene pipe, the cable was much thicker than I had expected, a 4 core armored monster! Maybe it was damaged, but not going to pull it out and redo it in a rush. Will check both ends. Any other tests we can perform to confirm it is the cable that is the cause?

 

[AndyPRK]

If its 4 core maybe use can use a spare core that is not damaged

 

[phatboy]

Not sure why you would need SWA if it runs under the floor slab, I would assume it was at least 50mm below the floor surface?

Could you drag it out, and replace with 10 or 16mm T+E? It wont be anywhere near as big, and be a lot more flexible.

Although it would have taken quite a bit to damage the SWA?!

 

[PrenticeBoyofDerry]

Could be the SWA was damaged before it was pulled through as I agree with above it would have took something to damage the SWA whilst pulling in.
But it could be a fault at the terminal ends, especially if they have got wet with the weather.

 

[FrogParty]

Could be the SWA was damaged before it was pulled through as I agree with above it would have took something to damage the SWA whilst pulling in. But it could be a fault at the terminal ends, especially if they have got wet with the weather.

The meter box got very wet some months ago, someone left it flapping open, but seems dry now and is enclosed in extra cladding. SWA does seem tough stuff to damage, no idea why the contractor insisted on using it, the duct is 400mm below floor level set into the substructure below the slab. The DNO supply cable is tiny by comparison and that is just in a trench.

The 4th core was attached to earth at one end. Disconnecting that and terminating did seem to improve things, then the tripping reappeared. Had wondered if the extra core was acting as an aerial picking up noise? If we can't find any damage at ends, could it be some mind of noise issue? Any kind of supression we can try? Rethreading the cable will be difficult.

 

[securespark]

Disconnect the buried cable temporarily.

Run a temporary supply of 16mm T&E across the floor. Connect it up in place of the suspect SWA and see if the tripping disappears. Then you will know for sure if the existing cable is the culprit or not. If it is, you (or someone) has to bite the bullet and replace it.

 

[FrogParty]

Disconnect the buried cable temporarily.

Run a temporary supply of 16mm T&E across the floor. Connect it up in place of the suspect SWA and see if the tripping disappears. Then you will know for sure if the existing cable is the culprit or not. If it is, you (or someone) has to bite the bullet and replace it.

It could come to that but such a temp supply would have to run in through a window as well as across a room and passage way. Also it can be days between trips, so it is difficult to know if the problem has been fixed, already had several false dawns.

Really need to be able to measure the nature of the trip somehow. If there was an earth leakage from one of the circuits then an RCD in the CU should trip first, not the RCBO with a time delay. So either there is an earth leakage from before the CU e.g. the SWA cable or ends are damaged in some way, or the RCBO is mistaking something for earth leakage. Any way we can test for the latter?

 

[FrogParty]

Just to update things I have had the insulation resistance of the buried cable checked, and the connections either end, all look fine. All kinds of checks on the internal wiring showed nothing unusual. I remain convinced that the problem is the other side of the CU, but the SWA does not seem to be the cause.

Starting from outside I contacted the DNO to ask if our supply could be checked, they admitted that the pole transformer (only serving us) could be at fault or the 50m of their cable running to the meter, and would send someone to record the voltages.

Really need to be able to measure the nature of the trip somehow

Seems this would be a Power Quality Digital Storage Oscilloscope, hopefully the DNO has one.

Meanwhile the electrician replaced the RCBO with yet another unit (different make this time). Perhaps that has fixed it, because we have not had a trip for 4 days now, fingers crossed....

Could it just be a faulty batch of RCBOs? Or has the DNO done something to our supply? I guess the test would be to plug the old RCBO back in and see what happens, but for now too excited about having continuous power.

 

[JohnD]

you say it's an RCBO. That's unusual in the circumstances you describe. why was it put adjacent to the incoming supply?

 

[FrogParty]

you say it's an RCBO. That's unusual in the circumstances you describe. why was it put adjacent to the incoming supply?

System is TT with a metal CU, I was told that 2A could flow in the case without blowing the DNO supplied fuse, hence the trip before the CU (also protecting the cable run as well). Time delayed for discrimination between the 30mA RCDs in the CU. Pretty sure it is an RCBO (not just and RCD), the overcurrent protection there to trip rather than blow the DNO supplied fuse if too much current. A bit out of my depth, but think I am accurately reporting what I have been told.

 

[JohnD]

if the main one is time-delayed and 100mA, and the 30mA ones in the CU have not tripped, surely we are looking either at multiple small leakages, or a fault in the cable to the CU.

What brand of CU is it?

Are there any non-RCD-protected circuits in the CU?

Have you had any water penetration?

 

[FrogParty]

Thanks for taking an interest John

if the main one is time-delayed and 100mA, and the 30mA ones in the CU have not tripped, surely we are looking either at multiple small leakages, or a fault in the cable to the CU.

Happened with just one RCD in the CU (that did not go) running a couple of sockets, so that suggests not multiple leaks. Checked supply cable thoroughly short of replacing it which would be difficult.

What brand of CU is it?

Sorry don't know, has a logo of black disc with red triangle pointing out of it.

Are there any non-RCD-protected circuits in the CU?

No.

Have you had any water penetration?

See above, new wet plaster, rain in meter box (dried out).

So difficult to prove a negative, but changing brand of trip seems to have helped. Without a PQ-DSO I can'rprove it but I am convinced this is a noise problem rather than earth leakage.